Article Text

Is melatonin likely to help children with neurodevelopmental disability and chronic severe sleep problems?
  1. Caroline Willey1,
  2. Bob Phillips2
  1. 1Staff Grade Community Paediatrician, Central Manchester PCT, UK
  2. 2Associate Fellow, Centre for Evidence-based Medicine, Oxford, UK

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

A girl aged 3 years 6 months has neurofibromatosis with significant visual impairment and mild to moderate learning difficulties. She has always been difficult to settle to sleep and has frequent nocturnal wakenings.

A sleep programme with specific behavioural management techniques has been used, as have sedative medications such as trimeprazine, which caused deterioration in concentration and daytime sleepiness. Should she be tried on melatonin?

Structured clinical question

In a preschool child with visual impairment and mild to moderate learning difficulties, in whom conventional treatments have failed [patient] is melatonin [intervention] likely to improve her sleep pattern [outcome]?

Search strategy and outcome

Secondary sources

DARE, Clinical Evidence Dec 2000, Medicines for Children RCPCH 1999—none.

Cochrane Library: Systematic Reviews—0, Abstracts of Reviews of Effectiveness—0, Controlled Trials Register—six papers of which two relevant (same papers found through search detailed below).

Primary sources: Medline 1966 to present—using OVID interface AND Sleep (and exploded) AND Learning disabilities (and exploded) AND.limit to:Children <0 to 18 years> Human, English language.

This gave 90 references; all titles checked—15 considered, six included. Nine excluded as three non-systematic reviews, two other conditions, one non-delayed children, one slow release melatonin, two abstracts only.

Embase 1980 to present: searched with same strategy—no additional papers.

See table 2 for summary of the six included papers.

Table 2

Melatonin in sleep disorders


Most studies had small numbers of participants with significant drop out rates or non-randomisation in larger studies. Very few of the studies give p values or confidence intervals—they appear far too small to give statistically meaningful effects. One of the trials (Camfield et al) is very different in design, an “N-of-1” study. These trials are designed for each individual patient, and allow for interpersonal variation in drug effect. Classically an N-of-1 trial has three blocks; during each block the patient receives sequentially therapy and placebo under double blind conditions with an appropriate washout period. Response in two or three blocks is considered positive; less than this, caused by chance alone.

Even allowing for the difficulty of recruitment and objective assessment of outcomes in children with multiple difficulties, there is currently little good quality evidence for the effectiveness of melatonin. The startling increase in seizures noted by the Sheldon paper is of great concern, especially in the UK where melatonin is often given in an uncontrolled way with overseas imports of the drug. A large multicentre placebo controlled randomised controlled trial is needed to try to clarify which children and what types of sleep disorder are most amenable to treatment, and to define the likely side effect profile.

Clinical bottom line

  • Melatonin may be effective in sleep onset difficulties, but not in fragmented sleep or early morning wakening, though evidence is poor quality.

  • There is little evidence regarding melatonin’s long term safety profile.

  • Melatonin should be used with caution in any child with epilepsy in view of increased seizure frequency in one study; “N-of-1” methods may be considered.


Supplementary materials




    Table 2 Melatonin in sleep disorders

    CitationStudy groupLevel of evidenceOutcomesKey resultsComments

    Jan et al (1994)15 children aged 6 mth to 14 y, mean 6.8 y Most multiply disabled; 5 with epilepsy, 9 visually impaired Melatonin 2.5–5 mgDouble blind placebo controlled trial (level 1b); crossover studySleep charts Parental interviewNo adverse effects No response in 2/15 1 child—ceased effect even with 20 mg after 6 mth6 (40%) not randomised Type of sleep disturbance described
    O’Callaghan et al (1999)7 children aged 2–28 y with tuberose sclerosis with epilepsy + SLD Randomised to placebo or 5 mg melatonin 20 min prior to bedtime Crossover randomised double blind trial (level 1b)Sleep diary Total sleep time Sleep onset latency No. awakeningsMean improvement in total sleep time of 0.55 h (CI 0.088 to 1.01) No effect on fragmented sleep Sleep onset latency improved but did not reach significance Short treatment time for any adverse effects to become apparent No effect noted on seizure frequency
    Dodge and Wilson (2001)20 children with moderate to severe developmental disabilities (4/20 visual impairment); age range 1–12 years 36 recruited but only 20 completed study Randomised double blind placebo controlled trial (level 2b)Sleep latency Duration of sleep No. awakenings Sleep log and parental questionnaireSleep latency improved in all but 2 children on MLT (p<0.05); more marked in those with greater sleep latency on baseline measure Duration of sleep improved with MLT but no different from placebo No change in number of wakenings No side effects reported Large drop out rate but no reported differences in diagnosis, age, epilepsy, etc in those not completing No baseline data for type or severity of sleep problems in those dropping out
    Camfield et al (1996)6 children aged 3–13 y Blind with at least moderate learning disability, using 0.5–1 mg melatonin"N-of-1" double blind placebo trial (level 2b)Sleep diary Average numbers hours sleep per 24 hours Number of wakenings between 9 pm and 7 am Number of nights without wakening between 10 pm and 7 am Found MLT to be ineffective in 5/6Low dose used Timing in relation to desired sleep time may have been too long No adverse effects noted No information about blinding or randomisation
    Palm et al (1997)8 aged 3–23 y (6 children aged 18 or less) All functionally blind, M/S learning disabilities 0.5–2 mg melatonin, age dependentOpen study (level 4)Sleep diaries for 6 weeks prior to treatment and several months during treatment MLT levels in 7 children"Dramatic" response in all 8 Loss of effect in 1 after 6–8 months MLT levels showed delayed peakNo side effects reported
    Sheldon (1998) Published as research letter6 children, 9 months to 18 y Multiple neurological deficits and chronic sleep disorders with 5 mg at bedtimeOpen study, consecutive recruitment (level 4)Wrist actigraph Changes in sleep onset latency, nocturnal wakenings Total sleep timeMarked improvement in all 3 measures in 5/6Study stopped due to increased or new seizure type activity on melatonin in 4/6 No info on types of AE meds used

    MLT, melatonin.

  • .

    Archimedes Update

    The update is available as a downloadable PDF (printer friendly file).

    If you do not have Adobe Reader installed on your computer,
    you can download this free-of-charge, please Click here


    Files in this Data Supplement:

    • [View PDF] - Is Melatonin likely to help children with neurodevelopmental disability and chronic severe sleep problems?


  • Bob Phillips